This project is concerned with aspects of sentence comprehension. Its overall aim is to understand how listeners decode linguistic information and store it in memory. Most of the work is concerned with devising and testing process models for various aspects of comprehension. The project focusses upon the following issues: (1) Speech decoding - what codes are computed by listeners as they hear spoken speech? Related to this is the attempt to model the way in which subjects carry out an experimental task known as phoneme monitoring. (2) Context effects on lexical access - this part of the project is concerned with determining how listeners use semantic and syntactic information to aid in retrieving words in sentences from the "mental dictionary." (3) Studies testing some general models of sentence memory. (4) The effects of text structure on sentence comprehension. In this work we measure reading times for sentences as a function of the role that these sentences play in the text as a whole.